摘要
根据含噪语音中噪声、清音和浊音的不同特点,提出了一种改进的多尺度小波域阈值去噪方法。采用常用的软阈值函数对清音和浊音进行不同的阈值处理,而对于噪声,将其对应的小波系数全部置零。该方法在很大程度上抑制了噪声,又减少了语音段清音的损失,提高了信噪比。仿真结果表明了该方法的可行性和有效性。
According to the differences between noise, sonants and surds in the noisy speech, a modified speech denoising algorithm using threshold-based wavelet in wavelet domain method is proposed. Different processing methods are used to restrain the noise from sonants and surds. Simulations show the effectiveness and feasibility to reduce the loss of surds and to enhance the SNR（Signal-to-noise Ratio）.
出处
《电声技术》
2007年第12期51-53,60,共4页
Audio Engineering
关键词
小波变换
闽值
峭度
wavelet transformation
threshold
kurtosis
